scientific article; zbMATH DE number 4022301

From MaRDI portal
Publication:3764980





For a double sequence (\(\eta\) (n,m): \(n,m\in Z_+)\) of random variables valued in E, a discrete state space, one kind of Markov property is defined as: for \(n\geq u\) and \(m\geq v\) in \(Z_+\), the distribution of \(\eta\) (n,m) conditioned on \(F^*(u,v)\) (\(\equiv\) the \(\sigma\)-algebra generated by \(\eta\) (s,t) with either \(s\leq v\) or \(t\leq v)\) is the distribution of \(\eta\) (n,m) conditioned by: \(\eta\) (u,v), \(\eta\) (u,m) and \(\eta\) (n,v). It is proved that a necessary condition is that the associated sequence (\({\tilde \eta}_ n:\) \(n\in Z_+)\) of \(E^ Z\)-valued random variables is Markov. Here \({\tilde \eta}_ n(K)\equiv \eta (u,v)\), with \(n=\min (n,v)\) and \(K=u-v\in Z\). When \(\eta\) has a stationary transition, then \({\tilde \eta}\) too will have a stationary transition.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3764980)